In elevator control systems, besides the data on the load and calls, it is necessary to have data on the number of people in the elevator car in different situations during use. The numbers of passengers can be used, on the one hand, to produce statistics to monitor and enhance the control of the elevator and, on the other hand, to provide current information on the loading of the elevator. Both short-term and long-term statistics are maintained. The numbers of passengers entering and leaving the elevator car are recorded in the statistics separately for each floor and direction as functions of time. The statistics are utilized in the control of the elevator or the associated external devices. In group control, the statistics are used to control the elevators in such a way that the prevailing traffic type and intensity as well as the estimated number of persons behind a call are taken into account. Different floors are served according to the need. It is also possible to consider the prevailing traffic situation in the control of the open times and closing speeds of the doors so that, and the degree of admission of passengers into the car can be optimized. The data on the number of passengers can be used to give better information to the customers about the prevailing situation and to control people's behaviour. The long-term statistics can be used in the development of the activities. In some cases it is also necessary to obtain information on the real traffic over a relatively long period, even 30 days. In prior art, the number of persons moving into and out of an elevator car has been determined using photocells to detect the movement of people or by measuring the load during stoppage. A photocell has a limited power of resolution in peak traffic conditions, especially if passengers move simultaneously in both directions through the doors. Procedures using the load data involve measuring the total load of the elevator at the instants of stopping and departure and the minimum load between those instants. The number of people entering and leaving the elevator car is calculated from these results using an average passenger weight. Thus, the procedure is based on the assumption that all passengers leaving the car get out of it before those entering the car get in, which is not always in keeping with the actual situation. Inaccuracies also result from the differences between the real and the standardized passenger weights.